Professional Experience

AI Researcher, AI Hub, Canada (AI/HUB)

2023 – Present

Highlights:

  • Crafted LLM reports for a roster of close to 10 clients, enhancing business outreach and solutions.
  • Demonstrated exceptional research acumen that garnered a $1.75 million grant from TAC for the company.
  • Engineered integrated banking, retail, and e-commerce solutions that significantly increased business profitability.
Course Recommendations Project

This project involved creating an intelligent course recommendation system for an e-commerce platform specializing in educational content. The goal was to enhance user experience by providing personalized course suggestions, thereby increasing engagement and sales.

Key Acheivements
  • Designed a recommendation system for an education sector client, achieving a 20% increase in lesson engagement and a 15% growth in new subscriptions, substantially elevating revenue.
  • This strategic implementation led to marked improvements in course enrollment and completion, contributing to a significant revenue boost for the company.

Technologies: Hybrid Models, Python, Reporting, Flask, HTML, JavaScript

Digital Human Avatars Project

Leading the development of lifelike digital human avatars in virtual reality, this project involved creating realistic and interactive AI-driven avatars for immersive experiences.

Key Achievements
  • Utilized advanced datasets to create highly realistic digital human models, enhancing the immersive experience in virtual reality.
  • Developed natural language and gesture recognition capabilities, enabling intuitive interactions with the avatars.

Technologies: Unity 3D, Machine Learning, AI, 3D Modeling

Key accomplishments
  • Developed a housing market predictive model with 85% accuracy, identifying key price influencers and aiding investor and policy decisions.

  • Analyzed supply-demand in major cities, finding mismatches that informed targeted housing development recommendations.

  • Discovered a direct correlation between economic indicators and housing prices, crucial for investment timing and policy planning.

Technologies: Linear Regression, Decision Trees, Random Forest, Python, Pandas, NumPy, Scikit-Learn, SQL, Data Visualization (Tableau/Power BI), Flask, HTML, JavaScript, RESTful API

This project involved optimizing SQL queries and automating SQL report generation processes using advanced Large Language Models (LLM). The goal was to streamline data retrieval and reporting procedures, enhancing efficiency and accuracy in data analysis tasks.

Key Achievements
  • Developed efficient SQL query optimization techniques, significantly reducing data retrieval times and improving database performance.
  • Automated complex SQL report generation, leveraging LLM to facilitate faster and more accurate data analysis and reporting.
  • Successfully integrated LLM with SQL databases to create a more dynamic and responsive data analysis framework.

Technologies: SQL, Large Language Models (LLM), Database Optimization, Automated Reporting

This project focused on automating the extraction of pertinent information from clinical reports using advanced text processing techniques, and enhancing Large Language Model (LLM) processes for improved data analysis and decision-making in healthcare.

Key Achievements
  • Implemented cutting-edge text extraction methodologies to efficiently parse and interpret clinical report data.
  • Enhanced LLM processes to facilitate more accurate and insightful data analysis, aiding in medical research and patient care decisions.
  • Successfully integrated automated systems, reducing manual workload and improving the speed and accuracy of clinical data processing.

Technologies: Natural Language Processing, Large Language Models, Data Extraction, Healthcare Informatics

This project involved creating an advisory report for automating product recommendations in retail through chatbot interactions. The project leveraged advanced AI technologies, including Large Language Models (LLM) and the Azure Bot Framework, to analyze customer queries and provide personalized product suggestions.

Key Achievements
  • Implemented AI-driven chatbots using LLM and Azure Bot Framework to effectively analyze and respond to customer interactions.
  • Automated product recommendation system which enhanced customer engagement and increased sales metrics.
  • Prepared an extensive report detailing the integration process, challenges, and benefits of using advanced AI in retail environments.

Technologies: AI, Large Language Models, Azure Bot Framework, Machine Learning, Data Analytics

Data Scientist, Tiger Analytics

2019 – 2023

Highlights:

  • Mentoring initiative for the team, fostering skill and team development.
  • Revamped retail demand forecasting, automating over 40 models for sales and workforce management in 450 stores.
  • Enhanced insurance fraud detection with a achieving a 30% referral acceptance rate in SIU.
  • Optimized insurance claims triage, leading to a 15% process efficiency gain with ensembled model solution.

Developed comprehensive labor forecasting models for a key CPG client, enhancing inventory and sales strategy planning.

Key Achievements
  • Constructed and automated labor forecasting models for 60 stores, successfully expanding the process to 450 stores for streamlined workforce management.
  • Created sophisticated forecasting models for transactions, gross sales, item quantities, and Limited Time Offer (LTO) items, optimizing product availability and sales opportunities.

Technologies: Prophet, Deep AR, SARIMAX, AWS Sagemaker, Python, Tableau, Reporting

Re-engineered real-time demand and response models for clothing and accessories in the retail sector, enhancing catalog management and market adaptability.

Key Achievements
  • Redefined modeling techniques using Sagemaker, developing over 40 classification and regression models to optimize catalog accuracy and customer targeting.

Technologies: Boosting classifiers, Regressors, PySpark, AWS Sagemaker, DevOps, Tableau

Developed an innovative fraud detection model for auto bodily injury claims, significantly enhancing fraud detection and improving operational efficiency in Special Investigation Units (SIU) and Claims Review Offices (CRO).

Key Achievements
  • Enhanced the insurance fraud detection system, effectively reducing false positives and increasing the acceptance rate of non-fraudulent claims by 30%.
  • Facilitated round table discussions to develop business rules, crucial for strengthening the insurance company's fraud mitigation strategies.

Technologies: Boosting Classifiers, Python, Tableau, Reporting, AWS EMR, Oracle SQL

Revolutionized underwriting processes by automating data pipeline workflows, significantly streamlining the analysis of large volumes of applicant data for more accurate and efficient risk assessment in health insurance underwriting.

Key Achievements
  • Achieved a 40% reduction in data processing time and a 30% increase in the speed of risk assessment, leading to a 25% improvement in overall underwriting efficiency.

Technologies: Python, Pandas, SQL, AWS Lambda, AWS S3, Logistic Regression, XG Boost Classifier

This project involved developing a predictive analytics model focusing on home loan defaults. By analyzing three years of data from 50,000 home loan accounts and identifying key predictors like credit score and debt-to-income ratio, we implemented logistic regression and random forest algorithms to effectively predict potential loan defaulters.

Key Achievements
  • Achieved 78% accuracy in predicting potential loan defaulters, leading to a 15% reduction in new bad loans, thereby bolstering the bank's risk mitigation and financial stability.

Technologies: Python, Pandas, SQL, AWS Lambda, AWS S3, Logistic Regression, XG Boost Classifier

The project aimed at automating and refining the process of liabilities risk assessment in auto insurance. We introduced an advanced, automated triage solution that facilitated a significant shift of claims from complex systems to more centralized services. This was achieved through a comprehensive analysis of patterns in Physical Damages, Insured Liabilities, and Subrogation Potential.

Key Achievements
  • Developed three sophisticated predictive models, integrated into an ensemble Rule-Based Classification (RBC) system, supported by a specialized dashboard to enhance data visualization and decision-making processes.

Technologies: Gradient Boosting Classifiers, Regressors, Python, Tableau, Reporting, AWS EMR, Oracle SQL

Tasked with retraining and enhancing the Auto Subrogation Suite, this project aimed at revolutionizing the auto insurance sector's approach to claim handling efficiency. The initiative involved a comprehensive update and management of several models including those for Subrogation Auto Collision, Medical Pay, and Early Intervention claims.

Key Achievements
  • Significantly improved model accuracy and claim identification by 10% through innovative rule modifications and model retraining.
  • Successfully established automated pipelines and advanced dashboards, thereby streamlining operations and enhancing data-driven decision-making.

Technologies: Gradient Boosting Classifiers, Python, Tableau, Reporting, AWS EMR, Oracle SQL

This project was centered around enhancing the Property Salvage Model, with a special focus on improving the model's performance in post-deployment scenarios. The role involved deep investigations into model failures in varied cases like fire claims, vandalism, and natural disasters, identifying root causes and rectifying them.

Key Achievements
  • Designed and implemented a robust model pipeline that significantly boosted post-deployment performance, especially in complex claim scenarios.
  • Created a comprehensive Tableau dashboard for more efficient and seamless data analysis, aiding in effective salvage recovery strategies.

Technologies: Gradient Boosting Classifiers, Regressors, Python, Tableau, Reporting, AWS EMR, Oracle SQL

Undertook a significant project to enhance the Property Subrogation Model, aiming to revolutionize claim recovery processes in property insurance. This included the development of an automated modeling pipeline to streamline operations and the integration of a Tableau dashboard for efficient data analysis.

Key Achievements
  • Successfully achieved a 10% increase in potential subrogation recoveries by refining thresholds based on detailed insights garnered from the production model.
  • Developed an advanced automated pipeline and integrated a comprehensive Tableau dashboard, significantly enhancing data analysis and operational efficiency.

Technologies: Gradient Boosting Classifiers, Regressors, Python, Tableau, Reporting, AWS EMR, Oracle SQL

Data Scientist, Open Access Technology India

2018 – 2019

Highlights:

  • Developed predictive models for an energy company, enhancing forecasting accuracy by 30% in energy demand and production.
  • Innovated energy forecasting algorithms, yielding a 25% improvement in resource allocation and grid management.
  • Created a conceptual AR application for an energy client, enhancing field experience with detailed guidance for seamless rectification of field operations.

This project focused on developing predictive models for an energy company, aiming to enhance the accuracy of forecasting in energy demand and production. The project utilized advanced data analytics and machine learning techniques to address challenges in energy forecasting.

Key Achievements
  • Enhanced forecasting accuracy by 30%, significantly improving predictions in energy demand and production.
  • Implemented sophisticated predictive modeling techniques to analyze historical energy consumption data and various influencing factors.

Technologies: Machine Learning, Python, Data Analytics

This innovative project involved creating a conceptual AR (Augmented Reality) application tailored for a client in the energy sector. The application was designed to enhance the field experience by providing detailed, interactive guidance, thereby facilitating seamless rectification and optimization of field operations.

Key Achievements
  • Developed an AR application that significantly improved field operation efficiency through immersive and interactive guidance.
  • Played a key role in integrating real-time data and analytics into the AR environment, enhancing decision-making processes in field operations.

Technologies: Augmented Reality, Interactive Design, Real-Time Data Integration

Software Engineer, Data Science, ProArch IT Solutions

2016 – 2018

Highlights:

  • Developed an innovative chatbot for administrator-level tasks, improving ticket resolution efficiency.
  • Created an AR application for a retail client, enhancing in-store customer experience and engagement.
  • Actively participated as part of the presales team, contributing to data science and VR proof of concept (POC) projects.
  • Engaged in high-level meetings, including stakeholders and CEO calls, providing valuable insights and technical expertise.
  • Worked extensively with diverse technologies including Angular, Python, Node.js, Unity 3D, and C# MVC Core.

This project involved the development of an innovative chatbot designed to efficiently handle administrator-level tasks, such as password resets, thereby streamlining ticket resolution processes and improving overall IT support efficiency.

Key Achievements
  • Successfully developed a chatbot that automated administrator-level tasks, significantly reducing the time and resources required for ticket resolutions.
  • Enhanced IT support operations by automating routine tasks, allowing the IT team to focus on more complex issues.

Technologies: Chatbot Development, AI, Automation

This project involved creating an augmented reality (AR) application tailored for a retail client. The application was designed to enhance the in-store customer experience by providing detailed, interactive guidance, helping customers navigate the store and engage with products in a seamless and immersive manner.

Key Achievements
  • Developed an interactive AR application that improved the in-store journey for customers, making it more engaging and informative.
  • Utilized cutting-edge AR technology to create a more dynamic and personalized shopping experience.

Technologies: Augmented Reality, Interactive Design, Retail Experience Enhancement

Software Engineer Practitioner, Data Science, Concentrix

2015 – 2016

Highlights:

  • Assisted the team in developing an automated solution for display advertisements, involving analysis and tagging of ads to streamline the segmentation process and eliminate irrelevant imagery.

This project involved assisting the team in developing an automated solution for managing display advertisements. The focus was on analyzing and tagging ads to streamline the segmentation process, thereby enhancing the effectiveness of digital marketing efforts and eliminating irrelevant imagery.

Key Achievements
  • Played a crucial role in the analysis and tagging of advertisements, leading to more efficient ad segmentation and targeting.
  • Contributed to the development of an automated system that reduced manual effort and improved the relevance and impact of display advertising campaigns.

Technologies: Digital Advertising, Data Analysis, Automation Tools